Summer 2026¶
Explore the new resources added to the Developer Guide in Summer 2026. Discover tools, call a Dataiku Agent from your own application, and learn how to share a static webapp.
The summer months are usually dedicated to updating existing tutorials and cleaning up outdated content, so you can continue to rely on relevant examples and instructions. This page highlights selected major additions to the Developer Guide; routine updates, bug fixes, and minor improvements are not included. For product release notes, see our documentation.
Tech blog¶
Explore Datapond, the Dataiku tech blog, to discover tools and techniques you can apply to your projects. Learn from practical examples and technical experiences shared by Dataiku’s teams across engineering, open source, and data science.
Start with the Developer toolbox section to try libraries and frameworks, understand how they work, and adapt code examples to your own workflows. Browse the other sections to explore implementation choices and lessons you can use in your own development work.
Building agents with CrewAI¶
Learn how to organize LLM-powered agents with the CrewAI framework. In this blog post, you will create your first Python agent, define roles and tasks, and assemble a crew of agents that work together. You will also learn how to connect your crew to tools hosted in Dataiku through an MCP server, and inspect execution traces to follow your agents’ actions and tool calls.
Read Add order to LLM chaos with frameworks: CrewAI on the Dataiku tech blog.
Using Marimo with Dataiku¶
Learn how to use Marimo’s reactive Python notebooks with Dataiku. When you change a cell, Marimo automatically reruns the cells that depend on it, and stores your notebooks as Python files you can track with Git. In this blog post, you will connect a local Marimo notebook to your Dataiku instance through the API, or configure a Code Studio to run Marimo within Dataiku. You will also find guidance on converting your existing Jupyter notebooks.
Read Using Marimo with Dataiku on the Dataiku tech blog.
Using a Dataiku Agent in an external application¶
Learn how to call a Dataiku Agent from your own Python application using the dataiku-api-client package.
In this tutorial, you will configure your connection through environment variables,
retrieve your project and agent identifiers, and send a query to your agent.
Use the complete code example as a starting point for your application.
Using Dataiku’s Python packages
Using a Dataiku Agent in an external application
Agents and Tools for Generative AI
Generative AI